ERP classification by sub

load data sub by pic

load(file = "C:/Users/andre/Documents/R/data/all_mas.Rdata")

tidy data

all_mas <- all_mas %>% filter(electrode != 65)

all_mas$Picture.number <- all_mas$Picture.number %>% as.factor()
all_mas$PictureCat <- all_mas$PictureCat %>% as.factor()
all_mas$CatNum <- all_mas$CatNum %>% as.factor()
all_mas$subid <- all_mas$subid %>% as.factor()
all_mas$electrode <- all_mas$electrode %>% as.factor()

all_mas %>% head()

Predict 2 categories

prediction_categories <- c("Erotica", "Bodies")
filtered_mas <- all_mas %>% filter(PictureCat %in% prediction_categories)

for sub i in sub id for each cat assign erps to one of 3 folds for each fold assign fold to test data train support vector on fPCA train find fPCA for test based on train predict test save correct percentage

in algorithm

make 3 equally sized sets of ERPs per sub

get fpca scores for training and test

train and test

do for other folds

for number of subjects

split ERPs in to 3 equally sized sets

for number of sets

pull one set for testing

train on other sets

test on final set

save correct %